A multi-model approach to stakeholder engagement in complex environmental problems

Abstract

We describe the different types of models we used as part of an effort to inform policy-making aiming at the management of the Ningaloo coast in the Gascoyne region, Western Australia. This provides an overview of how these models interact, the different roles they cover, how they fit into a full decision making process and what we learnt about the stakeholders involved in our project via their use. When modelling is explicitly used to address socio-ecological issues, the key determinant of success is whether the models, their results and recommendations are taken up by stakeholders; such uptake in turn depends on addressing stakeholders’ concerns, on engaging them in the project, on ensuring they feel ownership of the decision process at large, and that they understand and trust the modelling effort. This observation has guided our approach and has resulted in treating ‘building a model’ as the catalyst, rather than the final aim, of the process. In other words, extensive interactions in order to introduce, showcase, discuss and tune the model used for final decision making have represented both a requirement and an opportunity to ensure (i) model relevance, (ii) its acceptance, (iii) that all information available in the stakeholder team was accounted for and (iv) that stakeholders holding different levels of understanding of modelling, what it does and what it can provide to decision-making could develop an informed opinion on its use. To fulfil these roles we developed five broad classes of models: conceptual models, toy-models, singlesystem models, shuttle-models and a full-system model. In conceptual models the main drivers of a system are highlighted for subsequent representation as components of the full-system model. This usually results in a diagram summarising our understanding of how the system works. In toy-models a problem is simplified in such a way that only a handful of components are included. The purpose of these models is mostly educational: we want to understand how each component affects the problem and in order to achieve this, we temporarily renounce a satisfactory understanding of the overall problem. In single-system models we include a fairly detailed representation of a single component of the system (in our case recreational fishing and tourism); these models can be used to introduce stakeholders to modelling, provide temporary results from the study of a single activity, which will feed into the development of the final full-system model, or address sector-specific issues. In shuttle-models, we include the minimum number of processes we believe are crucial for a basic understanding of the overall problem. We know these models are still too simple for full system description, but they provide a sufficient understanding to enable us to contemplate, build and use the more complex models needed for full problem description. The term ‘shuttle’ refers to taking us from a minimum to a full description of the problem, a journey which is necessary both to developers in model definition and parameterisation and to stakeholders in the interpretation of the final full-system model results. Finally, the full-system model includes all information collected through the project and addresses all scenarios of stakeholders concern, and whose definition has been greatly eased by use of the ‘simpler’ models. As an example, a conceptual model may identify fishing and tourism as the main drivers of a region; a toymodel may describe how catches affect fish stocks; a single-system model may include the effect of gear, regulations and other processes affecting recreational fishing; a shuttle-model may include a simplified representation of the interaction between fishing, tourism, and infrastructure development on the overall health of the local ecosystem; this will gradually ‘take’ us to comprehend the ‘full’ model which may include tourism pressure, fish market values, climate effect, larger food-webs, etc.